DE2539206A1 - METHOD FOR AUTOMATIC ADJUSTMENT OF SEMI-CONDUCTOR DISCS - Google Patents

Description

Process for the automatic alignment of semiconductor wafers.

The invention relates to a method for the automatic adjustment of layered materials, in particular of Semiconductor wafers, according to the preamble of the patent claim

It is known to produce structures on semiconductor surfaces in that one is located on the semiconductor surface Photoresist layer is exposed in areas and then developed. In a further process step, the areas of the surface layer to be structured of the semiconductor surface that are not covered by the photoresist are etched away. The photoresist layer is exposed in certain areas. either by means of a contact copy or by means of a predictive copy. A contact copy is understood to mean the pressing of a photo mask and the stained semiconductor wafer against one another. The projection copy is a 1: 1 copy or a reduced projection using a high-resolution Lens generated. The 1:10 reduced single-chip projection has decisive advantages, such as the high one Resolving power of 500 lines / mm and the high adjustment accuracy of +0.2 / µm. Because of the elimination of the mechanical If the mask and semiconductor wafer are stressed, error-free copies can be achieved. Other advantages are the lesser one Testing effort for the mask control and the fact that the expansion or bending of the semiconductor wafer is only insignificant affect the accuracy of fit of successive structural levels. There is a disadvantage to this projection however, in that the area-wise exposure on the entire semiconductor surface of, for example, 50 mm in diameter is very costly is because this requires multiple adjustment and exposure operations.

VPA 75 E 7086 vP 17 BIa

709811/0403

ORIGINAL INSPlCTED

_ 2 -

The object of the present invention is accordingly to provide an economical method with the aid of which automatic adjustment and exposure is enabled ".

This object is achieved by a method, as already mentioned at the beginning, which is carried out by the method described in the characterizing part of the patent claim 1 is marked.

The adjustment method according to the invention reduces the Projection method advantageously economical made.

The adjustment method according to the invention is advantageously suitable not only for automatic adjustment during projection exposure, but also for contact exposure.

A slight change in the adjustment structure, as is the case, for example, in high-temperature processes, can advantageously occurs in an oxidizing atmosphere. This change occurs because im Area of the alignment structure silicon is converted into SiOp, so that the reflective Si / SiOp interface is somewhat in moves the semiconductor crystal into it.

In the following the invention will be explained with reference to the figures and the description explained in more detail.

1 shows a schematic representation of the formation of pyramid-shaped depressions through anisotropic etching,

FIG. 2 shows a section through a pyramid-shaped one shown in FIG Deepening,

Fig.3 shows the behavior of light, which by means of an objective onto the disc surface and into a pyramidal shape Deepening is projected,

Fig. 4 shows the plan view of a semiconductor wafer with the alignment spots according to the invention

VPA 75 E 7086 709811 / <^ 403

5 shows an enlarged, schematic representation of a

the adjustment spots shown in Fig. 4. 6 show the individual commands in a schematic representation

for automatic adjustment,
9 shows a schematic representation of the top view of an adjustment spot in the adjusted state,

Figure 10 show a schematic representation of the production of \ j inventive Justierflecken ·

Fig. 14 show a schematic representation of an additional device

u 15

^ to a standard step and repeat device.

The semiconductor wafer is denoted by 1 in FIG. For the production of the adjustment structures utilize the etching structures known per se, which are used in the anisotropic etching of monocrystalline silicon arise in alkaline etching. Such a pyramidal etched structure or depression is shown in FIG shown and denoted by 2. The resulting from anisotropic etching in the (iiO) surface of the semiconductor layer 1 Well 2 is pyramidal and has, as shown in Figure 1, four inclined (111) surfaces 22, which the sides of the Form I ^ yramids protruding into the semiconductor layer 1. ·

In Figure 2 is a section through such a pyramid-shaped Depth created by anisotropic etching is shown. There is between the (11i) surfaces and the (110) surface the semiconductor layer 1 at an angle of approximately 55 °.

The following considerations led to the invention: If, as shown in FIG. 3, light is applied by means of an objective projected onto the surface of the semiconductor wafer 1, light rays 3 which impinge on the (100) surface are in the lens reflects back. Rays 31, which hit the oblique (111) areas of. Wells 2 meet, however, are not in the lens, which has an aperture ratio of <1: 1.4, reflected back. These areas therefore appear dark.

VPA 75 E 7086 709811/0403

According to the invention, for automatic adjustment, as shown in FIG. 4, adjustment fields are provided on the semiconductor wafer 1 41 to 44 made. The adjustment fields 41 and 42 are used for rough adjustment. The reticle points to the appropriate Set up a bright square 51 or 52, which is the same size as the square area of the Jüstierfel £ er 41 and 42. The adjustment fields 43 and 44 are used for fine adjustment. The reticle has a pattern at the corresponding points 53 or 54 (in FIG. 14), which is bright in the areas in which there is no reflection from the pane and in the rest Areas is dark.

In Figure 5 are preferred dimensions of the alignment pattern structures specified. The number of small depressions 2 belonging to each alignment pattern can be, for example, between 5 x 5 and 20 x 20 lie. The congruent patterns 53 and 54 are colored dark at the points that lie between the depressions 2.

In the following, the automatic adjustment process will now be explained with reference to FIGS. 6 to 8. It is assumed that the disk 1 can be inserted so precisely with respect to the reticle that the corresponding adjustment areas, as in FIG shown overlap. This is through a pre-adjustment Well possible if the dimensions of the adjustment areas are e.g. 50 / µm «50 / µm to 100 / µm» 100 / µm. The automatic adjustment process then runs as follows:

As already described, there is a strong misalignment in the initial state. However, there is an overlap of the adjustment ranges of reticle and disk in front. The adjustment area 41 of the semiconductor wafer and the adjustment area thus overlap 51 of the reticle. Furthermore, the adjustment area 42 of the semiconductor wafer and the adjustment area 52 of the overlap Reticles. The first command for the automatic adjustment is now as follows: "Reticle in direction χ or -x so far move until two photocells, each one the whole, through

VPA 75 E 7086 709811 / 0A03

the light squares 51 and 52 register light which is passing through and is reflected back from the disk, indicating a minimum. " This means that as many of the light rays passing through the openings 51 and 52 of the reticle as possible then hit the Alignment spots 41 and 42 of the semiconductor wafer impinge and are deflected there in such a way that they do not pass through the openings 51 and 52 are reflected back, i.e. the assignment shown in FIG. 7 then lies between the alignment spots 41 and 42 and the openings 51 and 52 of the reticle. The second command required for adjustment is now: "Reticle so far Rotate (angle \ (J in Fig. 6) until the two photocells are the same Show value. "This state is reached when the number of light rays passing through the opening 51 of the reticle pass through and impinge on the alignment spot of the semiconductor wafer 1, is just as large as the number of light beams, which pass through the opening 52 of the reticle and impinge on the alignment spot 42 of the semiconductor wafer 1. The position shown in Fig. 7 is then present. The coarse adjustment is now carried out with the aid of the third command. the The accuracy of this coarse adjustment is + 1 / µm. The third command is: "Move reticle in direction y or -y so far, until both photocells show a minimum. "This state is reached when, as shown in Fig. 9, the Openings 51 baw. 52 of the reticle with the adjustment spots 41 or 42 of the semiconductor wafer 1 cover.

The same command sequence for fine adjustment is now carried out, in which the adjustment patterns 43 and 44 or 53 and 54 are adjusted to one another. Here, too, the adjustment is made to the minimum intensity. Correct with the x or y fine adjustment the position of the reticle for the intensity minimum of the adjustment patterns 53 and 43 does not match the position for 54 and 44, see above the middle position between these two locations is to be approached. This creates a difference in the distance between the adjustment patterns 43-53 and 44-54 balanced. The mutual position shown in FIG. 10 is then present in the adjusted state. The hatched areas mean the dark

7098 11/040 3

VPA 75 E 7086

Areas on the reticle, the solid lines the boundaries between the light and dark areas on the reticle for the alignment patterns 53 and 54 and the dashed lines the edges of the depressions 2 on the chip surface on the alignment patterns 43 and 44.

The coarse adjustment is advantageously carried out with green light, that only insignificantly converts the photoresist photochemically. This ensures that the photoresist is only used during fine adjustment, so only within the adjustment structures, photochemically converted and thus - in the case of positive resist - with Develop is removed.

The fine adjustment is advantageously carried out with the wavelength worked, which is used to generate the alignment pattern, since only with this wavelength an optimal coverage of the structures is achieved.

The adjustment structure suggested above is only applicable to silicon wafers possible with a crystallographic (100) surface. The adjustment structure proposed below is largely independent of the disk material and its crystallographic Orientation.

As shown in Figure 11, wet chemical or by means of ion etching, for example, 2 / µm deep holes 23 are etched into the substrate. With 231 the sloping slopes are the orientation (111) marked. In Fig.12 is the one shown in Fig.11 Section A-A shown. Since, particularly in the case of wet chemical etching, the course of the slope 231 at this stage is still fairly undefined, the substrate surface is bombarded with ions or atoms to create slopes around the etched flanks to be transferred with a clearly defined, uniform slope angle. The substrate 11 is not covered by a masking layer. In the case of silicon, for example, would be a SiOo layer on the silicon does not interfere. It is

VPA 75 E 7086 709811/0403

then stands the structure shown in section in Fig.13, the Has slopes with a flank angle of approx. 65 °. The manufacture of such angles of repose is from A. D.G. Stewart and M.W. Thompson in "Journal of Materials Science" 4, p.56 bis (1969). In Fig. 13 are the existing before the ion etching Embankments, as also in FIG. 12, are denoted by 231. The defined slopes that occur after ion etching are labeled 232. Just as in the case of the adjustment structure with the pyramid-shaped depressions described at the outset In the case of the adjustment structures according to FIG. 13, light beams which fall within the slope regions 233 do not ground into the objective reflected back so that the substrate surface in the area the embankment appears dark.

If the associated reticle adjustment structure is designed so that this is translucent at the points corresponding to the slope regions 233 and dark in the other regions, see above in turn, with the help of photocells, the light intensity incident on the photocells can be adjusted to zero fine adjustment. "

In Fig.14 is a device for performing the invention Procedure shown. Details of Fig.14, which have already been described in connection with the other figures have the corresponding reference symbols. The table of the step and repeat device on which the semiconductor wafer is arranged, is denoted by 6. Such a table is for example in the publication D.S. Alles et al, "The Step and Repeat Camera ", The Bell System Techn. Journ., Vol.49" N0V.I97O, No. 9, p.2145 .... 2177. The example The 1:10 reducing lens of the step and repeat device is marked 7. With 91 the reticle, with 912 the reticle structures, these are the dark areas on the reticle, labeled. The table on which the reticle rests and the in The x, y and iiJ directions are adjustable, is denoted by 92. The pivotable screen 93 is arranged below the table,

VPA 75 E 7086 709811/0403

_ 8th -

has openings 931 and 932 through which the light beams 71 and 73, which pass through the alignment spots 51 and 52 of the reticle, can be directed onto the objective 7 and ground. All other light rays are masked out by the pivotable screen 93. The reticle is preferably illuminated with convergent light 10 which originates from a condenser and which has, for example, the wavelength p> = 436 nm. Light beams 71 and 73, which are projected onto the semiconductor wafer 1 by the objective 7 in the manner shown in the figure, pass through the alignment spots 53 and 54, respectively. The light 72 or 74 back-projected from the semiconductor wafer 1, which passes through the aperture openings 931 or 932 and through the alignment pattern 51 or 52 on the reticle 91, is partially transmitted via the semitransparent mirror 81 or 82 via the lenses 83 or 85 steered to the photocells 84 and 86, respectively. A corresponding arrangement, which for the sake of simplicity, is not shown in the figure, is provided for the fine adjustment. For this purpose, corresponding openings 933, 934 are then provided in the diaphragm 93 and the adjustment patterns 53 and 54 already described above are provided in the reticle. In addition, another pair of mirrors 811 and 821, a pair of lenses 831, 851 and another pair of photocells 841, 861 are provided. The reference symbols for the elements used for fine adjustment are given in round brackets after the reference symbols for the elements for coarse adjustment.

In a further development of the invention, instead of the individual photocells 84, 86 or 841, 861 two photocells each are provided. In this case, in Yerstellmotoren also the same correct adjustment direction, which can be derived from the difference between the two cell currents, can be specified.

The diaphragm ■ 93, which is only at the location of the alignment pattern openings 931 or 932 etc. has, after the adjustment of the reticle 91 and the semiconductor wafer 1 has been completed, the possible adjustment accuracy is about + 0.2 / µm, swiveled away to the

YPA 75.E 7086 709811/0403

project entire reticle pattern onto the semiconductor wafer.

The semiconductor wafer is then automatically increased by one Raster unit moved further, so that the alignment and exposure of the next chip in the same way as described above can be made.

8 claims
15 figures

YPA 75 E 7086 70981 1/0403

Claims (8)

Claims ( 1y method for the automatic adjustment of layered materials, in particular of semiconductor wafers, characterized in that at least four adjustment spots (41 to 44) are applied to the surface of the layered material (1), v / obei these adjustment spots are depressions (2, 23) have that convergent light (10) from a condenser is sent through a transparent body (91) on which the mask to be adjusted is applied, the mask containing adjustment spots (51 Ms 52) which are created by displacement of the transparent body with the The mask applied thereon with the alignment spots (41 to 44) must be brought into congruence, so that a diaphragm (93) is provided which contains recesses (931 to 933) which ensure that only light rays passing through the alignment spots (51 to 54) occur, impinge on the surface of the semiconductor-shaped material (i), these rays being projected via an objective (7) onto the surface that the light rays incident on the surface of the layered material are reflected back via the objective (7), the openings (931 to 934) of the diaphragm (93) and via the adjustment spots (51 to 52), and that these rays are not reflected back, when they hit the depressions of the alignment spots (41 to 44) of the layered material, that the reflected beams (72, 74) are directed onto light-sensitive elements (84, 86, 841, 861), that these light-sensitive elements send signals to motors which can move the transparent body (91) with the mask applied in the x and y directions and can rotate the transparent body through an angle 0, the transparent support (91) "is oriented so that the photosensitive elements • Receive as little reflected light as possible. VPA 75 E 7086 709811/0403 2. The method according to claim 1, characterized in that that for rough adjustment of the transparent body (91) is inserted so precisely that there are alignment spots (51, 52) of the mask applied to the transparent body (91) and alignment spots (41, 42) formed on the surface of the layer-shaped body (1) are applied, partially overlap, which is achieved by a pre-adjustment that the transparent body (91) in the x direction or in the -x direction is shifted until the two light-sensitive elements (83, 86) that through the alignment spots (51, 52) register light reflected back from the material (1), indicating a minimum that the the transparent body (91) is rotated until the two light-sensitive elements (51, 52) show the same value, that the transparent body (91) is shifted in the y or -y direction until both are light-sensitive Elements (84, 86) show a minimum. 3. The method according to claim 1 or 2, characterized in that for fine adjustment on the mask of the transparent body (91) alignment patches (53, 54) are provided at the points at which the alignment patches (43, 44) have depressions, have light spots and the places lying between the light spots are colored dark that the transparent body (91) is shifted in the x or -x direction until two photosensitive Elements (841, 861) register a minimum of the radiation reflected back from the material (1) that the body (91) is rotated until the two light-sensitive Elements (841, 861) show the same value and that the body (91) in the y- or -y-direction so far is shifted until the two light-sensitive elements (841, 861) show a minimum. 4. The method according to claim 1, characterized in that that for rough adjustment on the mask of the body (91) VPA 75 E 7086 709811/0403 Alignment patches (51, 52) are applied by the image scale of the objective, the alignment spots (41 or 42) of the layered material (1) correspond. 5. The method according to any one of claims 1 to 4, characterized that the coarse adjustment with green Light occurs, which only insignificantly converts the photoresist photomechanically. 6. The method according to any one of claims 1 to 5j, characterized in that the adjustment patches (41 to 44) are generated on the layered material (1) when using silicon wafers that by anisotropic Etching in a manner known per se on the alignment spots (41 to 44), pyramidal etched structures or depressions generated on the (11O) surface of the semiconductor layer (1) will. 7. The method according to any one of claims 1 to 5, characterized in that wet chemical or means Ion etching about 2 / um deep holes (23) in the layered material (1) etched v / ground that to achieve clearly defined Etched flanks that with. the holes (23) provided material (1) is bombarded with ions or atoms, whereby Slopes with a defined flank angle are created. 8. The method according to any one of claims 1 to 7, characterized in that the rays reflected back (72, 74) via transparent inclined mirrors (81, 82, 811, 821) and via lenses (83, 85, 831, 851) onto the light-sensitive Elements (84, 86, 841, 861) are steered. VPA 75 E 7086 709811/0 4 03